absurdly-simple-benchmark

A simple comparison of the 'performance' of various programming languages

The Test

In psuedocode:

x = 1.00001
start timing
do (x = x * 1.00001) one hundred thousand times
end timing
print the value of x
print how long it took

Languages

• c
• go
• python
• ruby

Expectations

• c will be fastest
• go will be comparable to c but probably slower
• python and ruby will be comparable in speed, but much slower than c or go

Versions

TODO the versions of software used in this test should be listed here.

Running The Tests

g++ timing.c -o timing && ./timing

go run timing.go

python timing.py

ruby timing.rb

Results

Really flakey numbers here (which I will improve later) I found roughly

Timing

Language	Time to run (ms)
c	0.7
go	1.5
python	50
ruby	120

Output

Language	Output
c	2.718260
go	2.718295419874667
python	2.71829541987
ruby	2.718322602828866

Thoughts

Mostly my anticipations were correct about the performance.

I would have expected go to be closer to c, and python to be closer to ruby.

I was surprised at the variability in the output values, even knowing how crazy floats can be.

Any thoughts on this would be most welcome.

I will be running it with many runs and min | avg | max values, as well as over a longer period of time.

I would prefer to run for a constant time and see how many iterations were possible, but deciding when to stop would almost certainly be the most expensive part of that test, so I think the existing method is better despite it taking orders of magnitude longer for the dynamic languages.